Extending the LOINC conceptual schema to support standardized assessment instruments.

Abstract

To extend the Clinical LOINC (Logical Observation Identifiers, Names, and Codes) semantic schema to support (1) the representation of common types of assessment instruments and (2) the disambiguation of versions and variants that may have differing reliability and validity. Psychometric theory and survey research framework, plus an existing tool for implementing many types of assessment instruments (Dialogix), were used to identify and model the attributes of instruments that affect reliability and validity. Four modifications to the LOINC semantic schema were proposed as a means for completely identifying, disambiguating, and operationalizing a broad range of assessment instruments. Assess the feasibility of modeling these attributes within LOINC, with and without the proposed extensions. The existing LOINC schema for supporting assessment instruments was unable to consistently meet either objective. In contrast, the proposed extensions were able to meet both objectives, because they are derived from the Dialogix schema, which already performs those tasks. These extensions to LOINC can facilitate the use, analysis, and improvement of assessment instruments and thereby may improve the detection and management of errors.